A Visual Analytics Approach for Traffic Flow Prediction Ensembles

Abstract

Traffic flow prediction plays a significant role in Intelligent Transportation Systems (ITS). Due to the variety of prediction models, the prediction results form an intricate structure of ensembles and hence leave a challenge of understanding and evaluating the ensembles from different perspectives. In this paper, we propose a novel visual analytics approach for analyzing the predicted ensembles. Our approach models the uncertainty of different traffic flow prediction results. The variations of space, time, and network structures of those results are presented with the visualization designs. The visual interface provides a suite of interactions to enhance exploration of the ensembles. With the system, analysts can discover some intrinsic patterns in the ensemble. We use real-world urban traffic data to demonstrate the effectiveness of our system.